video_to_sequence
video_to_sequence copied to clipboard
about Neural caption generator and show and tell using tensorflow
Dear jazz,
very good work. I have seen your both applications based on tensor flow and caffe. I am a student and i am very beginner. i installed tensor flow bazel using pip and it is working fine. could you please explain how you installed caffe and other required packages for running both applications.
Thanking you,
Hi. Thank you for your interest in my projects.
Basically you need bunch of science/math related Python packages including Numpy and Pandas. I highly recommend you to use "Anaconda Python" https://www.continuum.io/downloads
This installs most of the packages you may need.
You also may need OpenCV and Python-YouTube package OpenCV: http://opencv.org/ PyTube: https://github.com/nficano/pytube
After all, you need to install Caffe (https://github.com/BVLC/caffe)
- installing OpenCV and Caffe could be difficult... it took me for a while as well...
And about TensorFlow and Caffe, they might not run at the same time. In my codes, they do not really run simultaneously. Once it extracts features using Caffe first, and then use TensorFlow to train the NeuralNet.
Let me know if you have difficulties running the codes. Thank you.
-Taeksoo